In 2004, 681 shifts (5448 hours) of beam were scheduled, out of which 5295 hours of beam were effectively delivered (including 41 hours of refills). This represents a beam availability of 97.95 %. Dead time due to failures accounts for the remaining 2.05 %. It should be recalled that since February 2003, and throughout 2004, the current in the ring has been topped-up with "Front Ends open" in all filling modes. As a result, the refill time is considered as 'working USM time' and is included in the availability figures of the accelerator. The availability figure for 2004 is well in line with that of 2003, itself already a record. All runs but one were especially good with more than 98% availability! Run 2 was spoilt by several storms and electrical drops coming from the electricity supplier, which were not compensated for because the High Quality Power Supply (HQPS) was out of service at that time.

The mean time between failures reached the satisfactory score of 49.1 hours, which is well above the 2003 score. As in 2003, all failures occurring during a refill are taken into account. This was not the case in 2002 and provides an explanation for the 'artificial' drop of the MTBF between 2002 (58 hours) and 2004. See Table 4 for detailed statistics.

Of great importance for the beamlines, many long periods of delivery took place without a single failure. There were 22 periods greater than 3 days of uninterrupted beam The absolute record of beam delivery without a single failure occurred during run 2004-05 when the beam was delivered for 12 days with no interruption (Figure 157).


Fig. 157: Main global parameters of the electron beam.


In view of the aging of some components, and in spite of intensive preventive maintenance, it appears difficult to further improve the availability beyond the present figure and our goal is to maintain it at this level.

Filling Patterns

The multibunch filling pattern, which represents 67% of the total user time, is mostly delivered in uniform mode with a lifetime of ~80 hours at 200 mA. Despite a slightly lower lifetime, the 2*1/3 fill is also used for experiments needing a non-uniform time structure. The remaining 33% is shared between the 16 bunch, the 4 bunch and the hybrid filling patterns (Figure 158 and Figure 159).


Fig. 158: Distribution of the 2004 user time in the various filling modes.



Fig. 159: Evolution of filling patterns over the last 3 years.


In 2004, the 4-bunch mode (4*10 mA) replaced the single bunch mode (15 mA). This new mode was proposed following the difficulties experienced in 2002-2003 with the latter. Whereas it has been possible in the past to deliver an intensity of 20 mA, this is no longer possible, a situation linked to the increasing number of low gap chambers installed in the ID straight sections. The maximum current is now limited by horizontal beam instabilities and incoherent tune shift, which saturate the injection process. With four equidistant bunches of 10 mA each, the beam parameters are now more stable during the decay. Such a time structure also allows a higher average current than the single bunch and is therefore more attractive for all users.

The hybrid mode was delivered with a new filling pattern developed in 2003. It consists of 24 groups of 8 bunches of current close to 1 mA and a single bunch of 4 mA corresponding to a total current of 200 mA (see Figure 160). A purity higher than 109 between the filled and the unfilled bucket has been routinely achieved.


Fig. 160: Hybrid mode filling patterns.